Note: Descriptions are shown in the official language in which they were submitted.
CA 02244720 l998-07-28
W O98/24176 - PCT~US97/20872
CEL~R G~ SENSrll~E A M PLlFnER WIlla
E~GH COMMON MODE SIGNAL REJl~CTION
l~lk;LD OF TEIE ~NVENTION
The present invention relates to charge sens;Live am--plifiers~ and in particular, to charge
S sensitive amplifiers with dirrerel"ial input t~rmin~l~ and high com~non mode signal rejection.
BACKGROUND OF THE INVENTION
Charge sensitive amplifiers play an important role in ;...~ systems due to the low
capacity ofthe subject signals in tenns of current, i.e., charge, and duty cycle. For ~mrle, in large
10 area, flat panel im~in~ systems, such as im~in~ systems for medical and document im~in~
applications based upon amorphous silicon, the image sensor is typically arranged as an array of
pixels, each of which consists of a photo sensiLive element and a thin film transistor ~TFT).
In order achieve im~in~ frame rates suitable for video processing and display, all gate and
data line connections for the sensor are brought out to the edge of the array for connection to an
15 off-array control circuit co~ row selection and charge sensing cile~ ly. As the resolution
of such an array increases, more pixels become nece~-y for each data line, with the result being
less signal capacity for each pixel. Further, as the number of pixels increases for a given active
sensor area, the total parasitic c~p~c.it~nce on each data line increases since the size of each TFT
is dictated by the rate at which the imaging i,L~,l"alion is read out, or sampled, from the array, and
20 thelerole, cannot generally be scaled linearly in accordance with the pixel pitch.
In fluoroscopic im~n~, the signal levels ~soc;~ted with the . ..; .; n. . . n dose rate may be as
low as 600 ele(i~rulls, while the parasitic c~r~it~nre of each data line is approx.,..~t~ly 100
picofarads. Under these circ~lm~t~ncç~, the difficulty for any charge sensitive amplifier cQ~ d
-
CA 02244720 1998-07-28
WO98/24176 - PCT~US97/20872
to such a data line is to lirnit both the ltf and thermal noise components of the amplifier and the
input-referred power supply noise to a value less than one n~icrovolt.
~ conventional strategy for limiting charge sensitive amplifier noise is to use a single sided
ar~ cl~lre with a large p-channel metal oxide semicond~l~tQr field effect transistor (P-MOSF~T)
as the input device. This single sided ar~hi~eetllre can reduce the amrtifier noise by a factor as great
as the square root of t~,vo.
For many applications, the biggest noise problem related to the power supply involves
power supply noise from on-chip digital switching circuits. In large area im~ging, the pixels require
biasing which is coupled, at least capacitively if not directly, to the data lines. Hence, the noise on
10 the power supplies which bias the array is directly coupled to the input of the charge sensitive
~mrlifiers which are connected to the data lines. Since limiting the power supply noise to a value
of one microvolt l~MS is a very difficult tasl$ it is necess~ry to somehow reiect the power supply
noise inyected at the input. Unfortunately, the ideal structure for rejecting this noise is a .li~elellLal
input ~mp1ifi~r which requires a second P-MOSFET, thereby increasing the ~mrlifi~r noise across
the parasitic data line ç~p~cit~nce by a factor of at least the square root of two.
Accordingly, it would be desirable to have a charge sensitive amplifier which is capable of
rejecting power supply noise without increasing its own amplifier noise.
SUMMARY OF l~ NTION
A charge sensiLive amplifier in accordance with the present invention has high common
mode signal rejection for rejecting input referred noise, as well as high power supply noise
rejection Such a charge sensitive ~mp~ifier can be used advantageously in mllltiple channel charge
-
CA 02244720 1998-07-28
WO 98/24176 . - PCT~US97/20872
_
sensitive amplifier chips for reading out im~gin~ data from large area image sensors, such as
im~qgir~ systems for medical and document imaging applications based upon amorphous silicon.
A charge sensiLivt; amplifier with high commc-n mode signal rejection in accordance with
one embodiment of the present invention in~ es a current controlled transistor circuit, a voltage
5 controlled transistor circuit and an output circuit. The current controlled transistor circuit is
configured to receive a dc input lere.encc with a common mode input signal and in accordance
therewith provide a first common mode signal component. The ratio of the first common mode
signal component to the common mode input signal equals a first signal gain. The voltage
controlled transistor circuit is coupled to the current controlled transistor circuit and is configured
10 to receive a data input signal with the co-lu..on mode input signal and in accordance thelewiLIl
provide a data signal component and a second c-~.. on mode signal component. The ratio of the
second co~ lloll mode signal component to the COIl~ Oll mode input signal equals a second signal
gain. The output circuit is coupled to one or both of the current and voltage controlled transistor
circuits and is confi~lred to receive and combine the data signal component and the first and
15 second co-.c n mode signal co.,.l)ollents and in accordance therewith provide an output signal
which co--t;~onds to the data input signal. The ratio of the output signal to the data input signal
s~lbst~nti~lly equats the second signal gain. The first and second signal gains are s~ sl ~ Ally equal
and of opposite polarities and the first and second common mode signal components substantially
cancel one another.
In a l)r~r~ ;;d embodiment of the present invention, the current and voltage controlled
l~nsi~lor circuits are coupled together in a totem pole circuit configuration and have
semicon~ ctor output regions of opposite conductivity types (e.g., an NPN bipolar junction
transistor ~3rr) and a P-channel metal oxide sernicon-iuctor field effect transistor CMOSF~T),
CA 02244720 1998-07-28
-
W O98/24176 - PCT~US97/20872
respectively, or a PNP BJI and an N-channei MOSFET, respectively). In an alternative
embodiment, the current and voltage controlled transistor circuits are coupled together in an
opposing circuit configuration and have semiconductor output regions of the same conductivity
type (e.g., an NPN BJT and an N-channel MOSFET, respectively, or a PNP BJI and a P-channel
S MOSFET, respectively).
These and other features and advantages of the present invention w-vill be understood upon
coneid~ration ofthe following detailed description ofthe invention and the acco".p~..ying d,~v~i.~s.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 is a simplifi~l 5~ ;c diagram of a charge sensitive amplifier in accoldal1ce with
one embodiment of the present invention.
Figure 2 is a simplified sçl~ ic diagram of an alternative embodiment of the charge
sensitive amplifier of Figure 1.
Figure 3 is a ~;mplified s~h~m~tic diagram of a charge sensltive ~mp1ifi~r in accordance with
15 another embodiment ofthe present i~ve~llion.
Figure 4 is a simplified sf h~m~fic diagram of an alternative embodiment of the charge
sensilive ~mrlifier of Figure 3.
Figure ~ is a detailed sf~h~m~tic diagram of an amplifier CO~ i,l;ilg the charge sensitive
amplifier circuit of Pigure 1.
DETAILED DESCl~IPTION OF TIIE FIGURES
Referring to Figure 1, a charge sensitive ~mplifi-or with high CCillllllOII mode signal rejection
in accordance with one embodiment of the present invention in~ an NPN bipolar jllnr,tion
CA 02244720 1998-07-28
W O98/24176 . - PCTAUS97/20872
l-~nsi~lor ~BJT) Q1, a P-channel metal oxide semiconductor field ef~ect transistor ~P-MOSFET)
Ml and a load reeict~nce Rload, hllelconlle~;Led in a totem pole circuit configuration as shown.
This di~le"lial amplifier confi~-ration has its "positive,l' i.e., noninverting, input terminal at the
base oftransistor Ql and its "negative," i.e., inve~ting, input at the gate of Ll~n~i~lol Ml. Using
5 a BJT at the nol~..vG, lhlg input advantageously ~ul~ ules a shot noise generator in the emitter
current of a BJI for the noise voltage generator a~oci~ted with a MOSFET gate. (It should be
noted that the charge sensitive amplifier with high common mode signal rejection is typically
followed by an inverting gain stage which would, in most applications, reside within the feedbarl~
loop. Hence, the lli,lvt;,li.l~" and llnoninvertingll ~esign~tions are used herein based upon the
10 assumption that such inverting gain stage is present.)
The base ~G"nillal of l.a"~i~lor Q1 receives the noninverting input signal ~N+, while the gate
of transistor Ml receives the inverting input signal IN-. When used in an im~ing system, the
noninverting input signal, IN+ is a positive reference voltage, while the inverting input signal IN-
is the input signal from a selected column of pixel elem~nt~ within the sensor array. With the15 it;~i~Live load Rload at the collector oftransistor Ql, the output signal OUT is an ampli
fied signal
with high conll.,on mode input signal rejection. This high common mode input signal rejection is
realized due to the subst~nti~lly equal signal gains from the inputs IN~, IN- to the output Ol~.
The ~ulJ~L~lially equal signal gains A+, A- from the inputs rN+, IN- to the output Oll'r can
be seen based upon the following analysis. From the illVel Lillg input IN- to the output OUT the
20 signal gain A- is the product of the gain from the ill~/t;l Lilg input IN- to the emitter of transistor Ql
ml11tirlied by the gain from the emitter of ~ .slor Q1 to the output OUT. Hence, the inverting
input signal gain A- is the gain (l,~nscol-lh~ n~e) g",Ml of transistor Ml as a source follower, with
the L~ co~ ct~nce of Lr~ls;;~Lor Ql as its load, multiplied by the gain (transcon~-lct~nce) gmQl
CA 02244720 1998-07-28
WO98/24176 . - PCT~US97/20872
ofthe common base amplifier formed by transistor Ql and the load l e C:~L~ e Rload. This can be
A~ = gm 1 + gQml gm RLO~.D)
;ssed as follows:
A gn~ ,1 ~ RLC~D
gm 1 + gm ¦
With respect to the nonillv~ g input signal IN+, the input stage is transistor Ql as a
commnn emitter ~mrlifi~r with its emitter degeneration equal to the ~ n.scon~uc~nre of ll ansi~o
A ~ RLOAD
gm 1
Ml. Accordingly, the non;ll~rwling input signal gain A+ can be e~.c~sed as follows:
A ~ gm 1 ~ gm 1 ~ R~QAD
gm 1 + gm 1
Based upon the roregoillg, it can be seen that, at least to a first order apploxi~ ;on~ the
inverting A- and noni--~elling A+ input signal gains are sub~lS.nli~lly equal. While this analysis is
10 a~pru~;n~le in the sense that a number ûf higher order effects have not been con~ red (such as
--rB
CA 02244720 l998-07-28
W O98/24176 . PCTr~S97/20872
_ _
back gate m~ tinn in ~ or Ml), circuit ~im~ tions using models for a 0.8 micron BiCMOS
process demonstrate gain m~t~hing to within 0.5%, i.e., 46dB of common mode signal rejection.
In addition to the high colnmon mode signal rejection provided, the circuit of Figure 1 also
exhibits large rejection of noise received via its power supply terminals VDD, VSS. Such rejection
5 is due to the fact that the impedances looking into the ampl;fier from power supply terminals VDD
and VSS <i.e., the load ,e, ~ ce Rload in series with the collector impedance oftransistor Ql, and
the drain i.~ cdal~ce of transistor Ml, respectively) is very high, thereby effectively i.~ol~ting the
input termmals IN+, IN- from mo~ ting signals, e.g., noise on the power supply terminals VDD,
VSS.
One characteristic ofthis circuit which, depending upon the particular application, may be
viewed as a disadvantage is the relatively large input offset voltage (appl~x;.~A~e~y 1.5 volts)
n.qcçss~ry bc;~weell the two input terminals IN~ . However, for many applications involving
charge sensitive ~mplifiers such as this, the nonînverting input signal lN+ is typically a DC r~e t nc~
voltage. Hence, such an input offset voltage requirement is not a problem.
Referring to Figure 2, an ~lt~rn~tive embodiment ofthe charge sensitive amplifier of Figure
1 uses a PNP BJI for transistor Ql and an N-channel MOSFET for transistor M1, with their
respective interconnection~ reversed accordingly. While exhibiting s~,lanLially the same
advantages as the charge sensitive amplifier of Figure 1, this embodiment does have one
disadvantage in that using an N-MOSFET as the signal input device results in a higher level of l/f
20 noise.
Referring to Figure 3, an alternative embodiment of a charge sensitive ~mplifiF~r in
accordance with the present invention uses NPN BlT Q1 and N-MOS~ET Ml in an opposing
circuit confi~-ration biased with a current source ISS and providing a di~t;l,~ial output signal
CA 02244720 1998-07-28
W O98/24176 - PCTrUS97t20872
OUT. While this embodiment exhibits sl~bst~nti~lly the same advantages as the charge sensitive
amptifier of Figure 1, one disa-lv~,l~,e is higher l/f noise due to the use of an N-MOS~ET as the
signal input device.
Referring to Figure 4, an ~ e embodiment of the charge sensitive amplifier of Figure
5 3 uses PNP BJT Ql and P-MOSFET Ml in an opposing circuit configuration biased with a current
source IDD and providing a ~ rel~..Lial output signal OUT. While this embodiment also exhibits
s~;,~ lly the same advantages as the charge se,.silivG amplifier of Figure 1, one disadvantage
is lower gain due to the use of a PNP BJI .
l~eferring to Figure 5, an amplifier circuit co~ .;..g a charge sensitive amplifier in
10 accordance with the present invention, spe~ific~lly the circuit of Figure 1, has NPN BJT Ql and
P-MOS~T Ml coupled between positive VDD and negative VSS power supply terminals. The
load resistance at the collector of transistor Ql is implemented in the form of resistive circuit
provided by P-MOSFET M4 and ~PN BJTs Q4 and Q5. The drain terminal of transistorMl is coupled to the negative power supply terminal VSS via the channel of P-MOSFET M10.
15 As noted above, the noninverting input AINB is a DC~ ~ere,e,lce voltage of app~ 1y 2.5 volts,
while the inverting input signal AIN is a data signal centered about a DC level of app,ox,~,-alely 1
volt.
Based upon the fcrGgoi"g, it can be seen that a charge sensitive ~mplifi~r in accordance
with the present invention has a number of advantages when used in charged integrating amplifier:
20 dirrelel lial input; low input referred thermal and l/f noise; high power supply reiection ratio; high
common mode rejection ratio; and a high impedance input terminal.
Various other mt~rlific~tions and alterations in the structure and method of operation of this
invention will be a~p&lGIl~ to those skilled in the art without departing from the scope and spirit of
!-rB
CA 02244720 1998-07-28
WO 98/24176 . - PCT~US97/20872
the invention. Although the invention has been described in connection with specific plGrellGd
embo-lim~nt~, it should be understood that the invention as claimed should not be unduly limited
to such specific embo~imRnte It is intRn(led that the following claims define the scope of the
present invention and that ~lu~;lules and me~hC~1e within the scope of these claims and their
S e~luiY~lc~ be covered thereby.
